Upload
impellotyrannis
View
224
Download
0
Embed Size (px)
Citation preview
8/2/2019 DNI Blackbook: Semantic Data Management -1
1/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
2/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Blackbook2 is a J2EE server-based data integration
framework
Relies on open standards to promote robustness andinteroperability JENA, JUNG, Lucene, JAAS, D2RQ
Based on semantic web technologies RDF, RDF Schema, OWL, SPARQL Vocabulary agnostic
Provides a default web application interface, SOAP andRESTful interfaces
Blackbook2 is PL3 Appendix E certified (PL3+)
Overview
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
3/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Architecture
Visualization Layer
Infrastructure Layer
Data Source Layer1 2 3
Infrastructure Layer
Data Source Layer
8/2/2019 DNI Blackbook: Semantic Data Management -1
4/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
fas
User Interface
Advanced Search
Workspace
A front-end Google-like userinterface allows analysts to easilyperform keyword and attributebased searches.
Advanced Search
Search
8/2/2019 DNI Blackbook: Semantic Data Management -1
5/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
fas
User Interface
Google-like Results
Network
Different ways to view the sameinformation. Network , for example,displays entities of different typesand their relationships to otherentities.
8/2/2019 DNI Blackbook: Semantic Data Management -1
6/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
fas
User Interface
Timeline
Multi-Panel
Different ways to view the sameinformation. Timeline , for example,displays entities chronologically
C SS // O O
8/2/2019 DNI Blackbook: Semantic Data Management -1
7/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
fas
User Interface
Google Map
Google Earth
Allows analysts to visualize geospatialcontent using Google-map and GoogleEarth.
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
8/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
fas
User Interface
Analyst Notebook
Mediawiki
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
9/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Architecture
1 2 3
Visualization Layer
Infrastructure Layer
Data Source Layer
UNCLASSIFIED//FOUOUNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
10/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
The Semantic Web is the next generation of the current
web in which computers can interpret the meaning of theweb content because of explicit semantics provided inmarkup.
Semantic Web
Unicode URI
XML + namespace + xmlschema
RDF + RDF Schema
Ontology
Logic
Proof
Trust
DigitalSignature
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
11/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Example 1: Inference
An analyst creates:
Buster
1) EntityBuster
Jennifer
2) Entity Jennifer
Nicholas
3) Entity Nicholas
An analyst makes the assertion:
mother-of
5) Jennifermother-ofNicholas
4) Busterhusband-ofJennifer
husband-of
Blackbook system can infer:
7) Nicholas child-ofJennifer
child-of
6) Jenniferwife-ofBuster
wife-of
9) Nicholas [step]child-ofBuster
[step]child-of
8) Buster[step]father-ofNicholas
[step]father-of
10) Busteris-genderMale
is-gender
Male
11) Jenniferis-genderFemale
Female
is-gender
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
12/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Example 2: Invalid Logic Assertion
Busterhusband-of
child-of
wife-of
[step]child-of
[step]father-of
is-gender
Male
An analyst creates:
1) EntityDon
Don
An analyst makes the assertion:
2) Jenniferbrother-ofDon
brother-of
Jennifer
mother-of
Female
is-gender
Nicholas
Blackbook system can infer:
3) Invalid Assertion(Gender conflict)
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
13/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Buster
CandyBill
Example 3: Constraints & same-as
Mark William
An analyst makes the assertion:
has-brother
1)Busterhas-brotherBill
has-sister
2) Busterhas-sisterCandy
has-son
3) Candy has-son Mark
has-uncle
4) Mark has-uncle William
An analyst applies the constraint:
5)Busterhas-only-one-brotherBill
has-only-one-brother
Blackbook system can infer:
6) William same-as Bill
same-assame-as 7) Bill same-as William
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
14/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Algorithms, Security, AKB
32
1
Algorithm plug-ins can be added
Security PL3+ / User Credentials CASPORT
Scattered CastlesDIAS
Analysts can store assertions into anAnalytic Knowledge Base (AKB)
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
15/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
fas
User Interface
Workflow
Yahoo Pipes
Workflow allow analysts to define theorder of tasks, configure algorithmparameters, and batch processesconcurrently
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
16/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
fas
User Interface
Workspace
Workflow and Workspace allow analysts to definethe order of tasks, store them in private folders and/orshare them publicly with colleagues.
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
17/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
Architecture
Visualization Layer
Infrastructure Layer
Data Source Layer1 2 3
Infrastructure Layer
Data Source Layer
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
18/29
UNCLASSIFIED//FOUO
UNCLASSIFIED//FOUO
1
2
3
A
B
C
D
E
F
G
H
I
Original Datasource
AKB
Analyst Knowledge Base
E
G
K
C
D
J
A
B
C
D
E
F
G
H
I
Composite Knowledge
K
J
Composite Knowledge
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
19/29UNCLASSIFIED//FOUO
U
S
S
T
U
S
T
U
U
U
S
S
U
T
S
U
AKB
U
U
U
S
S
U
T
S
U
Original Datasource Analyst Knowledge Base Composite Knowledge
T T
S S
Composite Knowledge with Security
U
S
T
Unclassified
Secret
Top Secret
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
20/29UNCLASSIFIED//FOUO
U
S
S
T
U
S
T
U
U
U
S
S
U
T
S
U
AKB
U
U
U
S
S
U
T
S
U
Original Datasource Analyst Knowledge Base Composite Knowledge
T T
S S
Composite Knowledge with Confidence
30%50%
100%10%
30%50%
100%
10%
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
21/29UNCLASSIFIED//FOUO
fas
User Interface
Relationship Manager Entity Manager
Ontology Import
Allows analysts to specify the relation-ship between two or more entities Allows analysts to create entities ofdifferent types, and modify attributes
Allows analysts to upload their ownontology
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
22/29UNCLASSIFIED//FOUO
Unstructured/Structured to RDF
Translation
Tran
slatio
n
Unstructured
Structured
RDFRDMS/XML
Unlike most applications, Blackbook performs queries on
data in RDF form, not relational form.
Sarin
Use ofAgent
Tokyo
Japan
Threatto Use
Unknown
Israel
Tel Aviv
Use ofAgent
NewYork
UnitedStates
urn:a911report:incident233
urn:sandia:organization65
urn:anubis:incident873
UNCLASSIFIED//FOUOUNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
23/29UNCLASSIFIED//FOUO
Datasource Connectivity
UNCLASSIFIED//FOUO
UnstructuredText
TransformNet-OWL
D2RQ
MAP
LUCENE
TTT
RDF
TransformScript
RDF
TransformScript
RDFXXX
XML
RDBMS
XX
X
XML
JENA
MAP
MAP
Client
Stub
MAP
1
2
3
4
5
6Web-
Services
RDBMS
MAP
MAP
Currently, there are six avenues to connect Blackbook with
data; 1-3 requires offline translation to RDF, 4-5 uses
real-time translation using D2RQ, 6 uses secure web-services.
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
24/29UNCLASSIFIED//FOUO
Blackbook and Alternate Stores
Hbase, Lucene,
Solr
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
25/29
UNCLASSIFIED//FOUO
Scalability using Hadoop
Scalable: Hadoop can reliably store and process petabytes.Economical: It distributes the data and processing across clusters of commonly availablecomputers. These clusters can number into the thousands of nodes.Efficient: By distributing the data, Hadoop can process it in parallel on the nodes where thedata is located. This makes it extremely rapid.Reliable: Hadoop automatically maintains multiple copies of data and automatically redeployscomputing tasks based on failures.
Hadoop implements MapReduce, using the Hadoop DistributedFile System (HDFS). MapReduce divides applications into manysmall blocks of work. HDFS creates multiple replicas of datablocks for reliability, placing them on compute nodes around thecluster. MapReduce can then process the data where it islocated.
UNCLASSIFIED//FOUO
http://hadoop.apache.org/8/2/2019 DNI Blackbook: Semantic Data Management -1
26/29
UNCLASSIFIED//FOUO
Blackbook and Wikis
Wikis
Wikis
2
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
27/29
UNCLASSIFIED//FOUO
Blackbook and Wikis
MySQL
Oracle
SOAP/REST
Browser
Wiki
ex) Intellipedia
Like browsers, Wikis can be a front-end toBlackbook. Wikis can also be a datasource. Wikiextensions can be utilized to enable Semantic and
Blackbook features.
Semantic extension
Blackbook extension
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
28/29
UNCLASSIFIED//FOUO
Revolutionize Technology Transfer
Research andDevelopment
BlackbookInfrastructure
T&E andDeployment
Research Government
Self-Addressed Stamped Envelope
A research product (red), such as a new and improvedalgorithm or visualization, can easily be transferred fromresearch to government using the Blackbookenvelope .
Improve Intelligence Analysis by Coordinated Exposition of MultipleData Sources Across Intelligence Community Agencies
UNCLASSIFIED//FOUO
8/2/2019 DNI Blackbook: Semantic Data Management -1
29/29
Relational vs. Graph-based Systems
1 2 3
Infrastructure Layer
Data Source Layer
Most Systems
Relati
onal
1 2 3
Infrastructure Layer
Data Source LayerGraph
Blackbook2 is a JEE server-based RDF processor thatprovides an asynchronous interface to back-end
datasources.